Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

The interpretability of landsat colour composite images for a geographical study of the northern coastal zone of west Java

Kardono Damojuwono and P. J. Wisnusudibyo
(paper presented by Aziz Poniman and Mimi Munami)

Introduction

The north coast of West Java is interesting because it has shown rapid changes during the last few decades. The coastal zone consists mostly of recent fluvial and marine deposits, with some volcanic rock in the northwestern part.

Satellite remote sensing has become more and more important as a technique for regional studies, especially for less accessible large areas where previously information was very scarce. Most of the Indonesian coastal zone belongs to this category. Many features of the coastal zone, especially the geographical elements, can be identified through LANDSAT colour composite images.

This paper discusses the geographical interpretation of the 1976 LANDSAT photos. The interpretation was based on: physical characteristics; features developed by human influence; features representing the interaction between physical and man-made processes; and identifiable features and processes which are useful for coastal zone management.

A similar study has been published by Suwahyowono (1979) on the north coast of East Java, around the mouth of Bengawan Solo River. The aim of his study was to determine the extent to which a rural Land Use Classification could be made using the images of LANDSAT colour composite MSS bands 4,5, and 7 for multistage and multispectral analysis. Hadisumarno (1977) describes the use of LANDSAT images for geomorphological mapping in the Palu area.

Materials and Methods

The quality of the images used was very good: clear, clean, and relatively free of cloud cover. Using the normal enlarger, the contrasts of tone, colour, texture, and resolution were easily delineated. The colour composites consisted of band 4 (green), band 5 (red), and band 7 (near infrared).

The methods used in this study were photomorphic analysis and field observation. The results are presented on a map; geographical features are identified by image interpretation and field observations are indicated by letters keyed to the discussion in this text. The main geomorphological units as represented on the map using the ITC system of geomorphological survey are scarp zones, barren lands, fluvial terraces, flood plains, deression deposits, and alluvial coastal plains.

The photos are made in mosaic, scale 1 :250,000, as a colour composite image, processed directly from a Computer Compatible Tape (CCT) with double-edge enhancement by the Earth Satellite Corporation, USA. This mosaic represents frames as follows:

In a colour composite, the objects in the image can be differentiated by using a colour indicator. Vegetation generally gives a red colour reflection, and water appears as light blue, dark blue, and black, depending on the depth. Unvegetated dry land gives a white colour reflection, whereas wet land gives a bluish colour. A mixture of vegetation, water, and land gives transitions from red to blue to white (Purwadhi 1978).

To support the interpretation, the following maps were used:

1. Vegetation map of Java and Madura, 1950, scale 1 :1 ,000,000
2. Geological map of Indonesia, 1965, scale 1 :2,000,000
3. Soil observation map of West Java, 1963, scale 1 :250,000.

There are six distinctive units of soil in the study area (see Table 1).

Results and Discussion

From the LANDSAT photo and the existing maps, some of the following geographical features, important for coastal zone management, were interpreted (Fig.1).

TABLE. 1. Units of Soil in the North Coast of West Java

Source: Soil Observation Map of West Java, 1963, scale 1:250,000.

1. Physical Features

1.1 Boundary between land and sea The boundary between land and sea is clear on the LANDSAT imagery because of the colour reflection difference. Sea water gives a dark blue colour reflection while land gives variable colour reflection, from red, blue, green, to white.

1.2 Water bodies, rivers, river mouths, and beaches These features can be traced from the colour difference, colour gradation, texture, site, pattern, and association of each one.

1.2.1 The sea

On the LANDSAT colour composite, the sea is dark blue in colour.

1.2.2 Lagoons

Lagoons are water bodies within an atoll or behind barrier reefs or islands (Desannetes 1977). They show blue colour reflection because the water here is relatively shallow. Although small, a lagoon is easily detected from the LANDSAT colour composite on a scale of 1:250,000 because of its association with reefs and islands, for example, the lagoons at Kepulauan Seribu.

1.2.3 River mouths

River mouths are easily detected because of the distinctive light blue colour of the turbid water where they meet the sea. Some examples are the mouths of the Cimanuk, Citarum, Cilamaya, and Cipunegara rivers.

1.2.4 Beach ridges

Beach ridges of sand, gravel, or cobbles are built up by storm waves on the backshore (Desannetes 1977). Beach ridges are seen as lines parallel to the shoreline, with a whitish colour reflection. An example is the beach ridge at Tanjung Bangkaderes near Cirebon.

1.2.5 Deltas

A delta is a deposit of sediments formed at the mouth of a river either in the ocean or lake (Desannetes 1977). Deltas are identified not by their colour reflection (because this depends on their land cover or land use) but rather by their site near the shore and its form. There are three types of deltas: bird-foot, cuspate, and arcuate. The bird-foot delta of the new Cimanuk River and the arcuate delta of the Citarum River are examples.

1.2.6 Alluvial fans

These are cone-shaped deposits of alluvium made by a stream where it runs out from high ground onto a level plain. A fan is generally formed where a stream flows from a mountain to the low land (Desannetes 1977). Therefore, alluvial fans, among others, can be detected from their site, with a river branching from a mountain front. In general, alluvial fans are fertile land and are therefore used for agriculture, such as rice fields or dry farming. In a LANDSAT colour composite such alluvial fans look like a triangle, violet to blue or pale blue, sometimes mixed with reddish colour. An example is the alluvial fans at the foot of Mt. Batur at the western end of the north coast of West Java.

2. Features Developed by Human Influence

2.1 Settlements and built-up areas (K)

An administrative criterion (population) is generally used to distinguish a town from a village (Malingreau, 1977). Such areas are blue and pale in colour reflection, associated with some linear patterns which contrast with the surounding colour. Some examples are the urban zones of Jakarta, Indramayu, Cirebon, Jatibarang, and Pamanukan.

2.2 Road network

The road network is a very clear linear pattern, in contrast to the surrounding colour features. Examples are the Jakarta-Cirebon and Jagorawi highways.

2.3 Irrigation network

This is characterized by its patterns of bluish lines and its association with rivers, river branches, or rice field areas, as in the irrigation networks around the Cimanuk, Cipunegara, and Citarum rivers.

2.4 Present land use and land cover

Land use is any type of permanent or cyclic human intervention to satisfy human needs, either material or spiritual or both, from the complex of natural and artificial resources. Present land use is by no means static (Malingreau 1 978). Data can be collected by ground survey (field observations, questionnaires) or by the interpretation of air photos and other remote imagery. When the latter method is adapted, land use is inferred from the study of the image characteristics, which are directly related to the land cover present at the time of the flight (Malingresu 1977).

In this interpretation, land use when the LANDSAT photos were taken can be deduced from present land cover features. Existing land use and land cover along the north coast of West Java that can be identified from LANDSAT Image colour composites on a scale of 1 :250,000 are as follows:

2.4.1 Irrigated rice fields or sawah (Si)

A sawah is a rice field which is artificially supplied with water (Malingreau 1977). Irrigated sawah features, whether in ploughing or in planting periods, give greenish blue or pale blue colour reflection due to the water cover. As the rice ripens it gives a bluish pink colour reflection. In the harvest period, or immediately after harvesting, the colour reflection is yellow. The greenish blue or pale colour reflection depends largely on the physical properties of each soil type. Irrigated sawah occur along the Cimanuk, Cipunegara, and Citarum rivers, and also to the southeast of Cirebon.

2.4.2 Rainfed sawah (St)

Rainfed sawah are sawah which are watered only by impounded rain water, sometimes supplemented by a very localized runoff collection system (Malingreau 1977). Rainfed sawah, whether in the ploughed period or the planting period, give greenish or pale colour reflection, depending on the physical properties of the soil type. They do not give a blue colour reflection like the irrigated sawah because the soil is not always water-covered. As the rice ripens it gives a pink colour reflection, and in or after the harvest period it gives a yellow colour reflection. An example of a rainfed sawah feature is seen to the west of Jatibarang (Losarang area).

2.4.3 Field crops (Tg)

Field crops show greenish or pale colouration with red spots which represent the soils and green vegetation. Such features are seen south of Jakarta as far as Bogor.

2.4.4 Tidal swamps

Tidal swamps are influenced by regular incursions of salty water and are usually covered with mangrove (Desannete 1977). A tidal swamp is seen as greenish blue with dark brown spots. Generally, it is located near a seashore or associated with river mouths, such as Muara Gembong, Muara Mati, Muara Pondok Tengah, and Muara Bekasi, which are located on the coast of Jakarta Bay.

2.4.5 Brackish fish ponds (T)

Brackish fish ponds are bodies of water managed for fish production (Malingreau 1977). They appear as bluish green with many dark brown spots. They are generally located near the coast, for example along Banten Bay to Pontang Cape, and in several places along the north coast.

2.4.6 Tidal forests (Hp)

These are brackish-water coastal forests which include mangrove (Avicennia, Rhizophora, etc.) and nipah (Malingreau 1977). A tidal forest gives a red to bluish colour reflection. Generally it is located on the coastal fringe, near the mouths of rivers as on Krawang Cape (in the area of the Citarum River mouth) and at Pamanukan Cape (in the area of the Cipunegara River mouth).

2.4.7 Teak forests (Hj)

In LANDSAT colour composites teak forest give a red colour reflection, sometimes mixed with brown and paler colours during the leaf-fall period in autumn. An example is the teak forest southwest of Jatibarang.

3. Features Representing Interaction Between Physical and Man-made Processes

3.1 Harbours A harbour appears as a short inlet with blue colour, located on the coast near a city. Tanjung Priok harbour in Jakarta is an example.

3.2. Drainage canals Artificial drainage systems can be traced on LANDSAT colour composite images. They are relatively more linear than the mother river, generally flow through rice fields or wet land agriculture areas, and originate from a reservoir. There are examples of canal systems along the Citarum, Cipunegara, and Cimanuk rivers.

3.3 Dams and reservoirs (L) Dams and reservoirs can be detected by their blue colour reflection. If the water bodies are turbid or shallow they become light blue, while clear and deep water becomes dark blue. Dam and reservoir building is usually associated with irrigation networks or with management of rivers. An example is the Jatiluhur reservoir.

3.4 Settlement patterns The pattern of linear settlement along a river looks like red spots, sometimes mixed with white and following the river. The red spots represent the reflection of karangkitri (garden) vegetation. A linear settlement pattern appears along the banks of the Ciujung, Citarum, and Cimanuk rivers.

3.5 Erosion ( Lk) Areas where erosion is very active are usually associated with barren land. These areas appear as brownish yellow reflections and usually are located in rough topography as around the Gede and Salak volcanoes in the west of the north coast of West Java.

4. Identifiable Features and Processes Useful for Coastal Zone Management

4.1 Water pollution or sea-water turbidity

Pollution or turbidity in sea water can be identified in a LANDSAT photo. Generally, water gives blue colour reflection with a variation from dark blue to light blue, and the cleaner and deeper water bodies tend to give dark blue. Pollution caused by sediments, hot water, petroleum, synthethic chemicals such as detergents and fertilizers, and organic wastes such as domestic and industrial wastes brought down by the rivers will influence the transparency of sea water and cause a difference in spectrum characteristics, e.g., from violet blue to greenish blue. Examples are the turbid coastal sea water in Jakarta Bay and in several places along the north coast of West Java, such as northeast of Cirebon.

4.2 Sedimentation in the waters around a river mouth

Sedimentation can be identified by its light blue to greenish blue reflection around the mouths of rivers and in coastal waters. This sedimentation is from the suspended load produced by erosion upstream and brought by the river and flood canals down to the coast. Sedimentation occurs in coastal waters from the mouth of the Citarum River to the mouth of the Cipunegara River and also from the mouth of the Cimanuk River to the coastal waters off Cirebon.

4.3 Excavation of clay for roof-tile and brick making (P1 )

The excavation of clay for roof-tile and brick making shows as pale brown in a speckled pattern around sawah areas, for instance in the Cikarang area.

4.4 Identification and monitoring of the extent of forest area and detection of illegal forest clearings Primary forest areas are characterized by a dark red colour reflection, and different textures depend upon the type and age of vegetation and the pattern of growth. Other land use or land cover types will also give different colours and textures. Thus a LANDSAT colour composite can be used for controlling illegal forest clearing or shifting cultivation in forest areas. Illegal forest cutting will be shown by the change from dark red to pale pink, while shifting cultivation will be seen as greenish or pale, and sometimes red, speckled with rough texture and located in the inner part of a forest.

4.5 Study of the accessibility of settlements along the coast

An estimation of the accessibility of a settlement along the coast can be obtained from the density of the transportation network in the area. An assumption can be made that the more roads leading to the settlement area, the more accessible the settlement, and vice versa. Settlements along the north coast of West Java based on the above assumption are considered less accessible, including those from the mouth of the Cimanuk River to Indramayu Cape, from the Anyar River to Cimanuk Cape, and from the mouth of the Cipunegara River to Pamanukan Cape.

4.6 Evaluation of the suitability of existing settlements on the coastal zone for future development The suitability of existing or proposed settlements can be evaluated by studying the environmental features in these areas. On the north coast of West Java, for example, the settlements around the north part of the Citarum River, around Pamanukan Cape, and around Cimanuk and Indramayu capes are in areas unsuitable for settlement, due in part to the many swamps and tidal forests.

Conclusions

Conclusions that can be drawn from this paper are:

1. Remote sensing techniques such as LANDSAT are adequate for studying large areas.
2. The colour reflection of the earth's surface in LANDSAT colour composite images depends on the enhancement processing and the colour composition of the bands used.
3. Field observation is necessary to confirm the interpretation results.
4. Even small objects can in some cases be identified from LANDSAT imagery, using a multistage surveying system and field observation.
5. The accuracy and speed of identifying objects depend largely on the image quality and the experience and capability of the interpreter.
6. LANDSAT colour composite images on a scale of 1:250,000 are very interpretable for a geographical study in the coastal area of West Java and are useful to guide coastal zone management.

References

Darmoyuwono, Kardono, 1975. Penerapan penginderean jauh di Indonesia, Lokakarya Evaluasi Pandidikan Penginderaan Jauh den Penerapannya di dalam Menunjang Pembangunan, Fak. Geograpi UGM, Jogyakarta.

_____________1976. Penerapan Penginderean Jauh dalam Mendetekti den rnemonitor Pencemaran Lingkungan Laut di Indonesia, Preceding Seminar Pencemaran Laut. LON-LIPI, Jakarta.

Desannetes V. R.1977. Catalogue of landforms for Indonesia. Soil Research Institute Bogor, Bogor.

Estes, J. E. oral. 1975. Fundamental. of image interpetation. In: Manual of remote sensing, R. G. Reeves, ed. American Society of Photogrammetry.

Hadisumarno, Surastopo, 1977. The geomorphology of Palu area Sulawesi from LANDSAT-I. The Indonesian Journal of Geography 7:34.

_____________1978. Metode penulisan laporan. PUSPICS UGMBAKOSURTANAL, Jogyakarta.

Malingresu, J. P., 1977. A proposed land cover/land use classfication and its use with remote sensing data in Indonesia. The Indonesian Journal of Geography 7 :33.

_____________1978. Rural land use image interpretation for its inventory and analysis. PUSPICS UGM-BAKOSURTANAL, Jogyakarta.

Purwadhi, Hardiyanti, 1978. Datar-dasar Interpretasi data data remote sensing. PUSPICS UGM-BAKOSURTANAL, Jogyakarte.

Sutanto 1978. Dasar-dasar interpretasi citra. PUSPICS UGMBAKOSURTANAL, Jogyakarta.

Suwahyuwono 1979. Klasifikasi penggunaan lahan pedesaan melalui citra Landsat komposit berwarna den foto udara merah infra di Kecamatan Bungah den sekitarnya, Laporan Akhir PUSPICSUGMBAKOSURTANAL Angkatan ke IV, Jogyakarta.

Discussion

Hehanussa: Do you have information on colour composites of LANDSAT I imagery of West Java other than those made in June 19767

Poniman: No composite was made of imageries obtained before and after June 1976, because the necessary imageries were not available, due to cloud cover.

Birowo: Why are certain area of Indonesia uncovered?

Poniman: They were covered by clouds of more than 60 per cent.

Siregar: Is it possible to detect enceng gondok (Eichornia crassipes)?

Poniman: It is possible to detect this plant by aerial photography, but not from satellite imageries. (Dr. Hehuwat explained that these plants were nevertheless detected at Rawa Pening using LANDSAT data.)

Siregar: How do you analyze erosion of riverbanks using LANDSAT data?

Poniman: It is possible to see undercut banks.

Sutamihardja: What is the percentage of dense forest in the Cimanuk area from analysis of LANDSAT imageries;

Poniman: It can be determined, but has not been done yet. From aerial photography it is about 40 per cent.

Ongkosongo: The US Topographic Army Map of 1972 might not be correct, because it was based on earlier data. According to available data at that time changes had occurred.

Poniman: I fully agree with that statement.

Ongkosongo: Is it possible to determine abrasion and accretion from turbid water patterns?

Poniman: To determine the occurrence of abrasion and accretion, you would have to compare the LANDSAT imagery with old maps.

Water-quality assessement of the cimanuk watershed

R. T. M. Sutamihardja and Herman Haeruman Js.

Data on water quality [physical' chemical, and biological) of the Cimanuk watershed were collected directly from the field on two occasions, namely in January and September 1978. To support this primary data, secondary data were also collected, including the geological, topographical, climatological data, and also data on demography and land use (including industrial use).

Determination of sampling sites was based on the following criteria:

1. altitude
2. land condition
3. main stream and sub-stream
4. distance from transportation facilities

The parameters of water quality were analyzed in the field laboratory in situ, and the rest of the data were analyzed in the laboratories of Bogor Agricultural University and DPMA Bandung.

The Cimanuk watershed covers an area of more than 400,000 ha and is situated between 107°25' and 108 25' east longitude, and between 6 10' and 7° 30' south latitude. The Cimanuk watershed area could physiographically and topographically be divided into three regions:

1. the upper region (consisting of the upper Cimanuk sub watershed, Cipeles sub-watershed, and Cilutung subwatershed)
2. the central region (central Cimanuk sub-watershed)
3. the lower region (lower Cimanuk watershed)

The upper region has a type A climate, the central region type B, and the lower region type C. In 1977 the population density in the upper region was 450.13 persons/km², whereas for the central region and lower region the figure was 1109.05 and 323.78 persons/km², respectively. With regard to year the total population of the Cimanuk watershed was 2,296,254 (in 1971) and 2,359,345 (in 1977) with an average population density of 551.36 persons/km² in 1971, and 556.51 persons/km² in 1977, resulting in a population rate of increase of around 2.75 per cent during the six-year period.

The water debit of the upper region during the rainy season was one to three times as high as the levels during the dry season, whereas in the lower region the same parameter during the rainy season was six to ten times that in the dry season. The temperature of water in the upper region ranged between 15° and 19°C, and in the lower region between 20° and 29°C. Flooding is still a serious problem in the Cimanuk watershed during the rainy season, in contrast to the problem of water deficiency during the dry season.

Water temperature of the Cimanuk streams was found to be naturally normal, and no signs of thermal pollution were detected which could endanger the lives of aquatic organisms. Suspended solids were the main physical pollutant, especially during the rainy season. The suspended solid content of the Cimanuk River during the rainy season was 20 to 60 times as high as the level during the dry season. The suspended solid content of the Cipeles River during the rainy season was found to be 35 to 180 times the level during the dry season. The same parameter for the Cikamiri and Cikeruh rivers during the rainy season was 60 times the level during the dry season. The suspended solid contents of 2055.4 - 3907.4 g/l during the rainy season and of 36.2 - 173.1 g/l during the dry season were an indication of a quite high level of soil erosion in the Cimanuk watershed.

The dissolved salt (indicated as salinity and conductivity) content and its SAR level showed that water from the streams of the Cimanuk watershed was still suitable for fishery, agricultural, and domestic use. This was also true of the pH level.

From the standpoint of alkalinity, the hardness and the calcium content, the water of the Cimanuk watershed may be classified as having a hardness of medium to high grade. The levels of these parameters, however, were still below the critical threshold for domestic, agricultural, fishery, and industrial use.

The dissolved oxygen content of the observed streams was found to be still within the feasibility range for aquatic organisms, especially fish. This was due to the lotic type of the observed water with its turbulency current, which made it possible for the oxygen from the air to diffuse into the water easily.

At the observation stations of Ciseureupan and Limbangan during the rainy season, the ammonia-N content of the streams of the Cimanuk watershed was found to be still within the feasibility range for fish. For drinking water and as a raw-water supply, however, the water should first be treated, e.g., by boiling.

The chlorine content of the observed water was found to be below the threshold level for fish, except at the observation stations of Samarang and Tolengas during the rainy season. This was possibly due to the chlorine-containing pesticides frequently used in agricultural areas, which were carried by the run-off during the rainy season. This presumption was based on the fact that the areas in the vicinity above the observation stations were used for dry-land agriculture and for intensive sawah culture as well.

The organic substance (BOD and TOM) content showed that the streams of the Cimanuk watershed already carried low to medium levels of organic pollutants.

The MPN-Coliform level showed that the streams of the Cimanuk watershed were already polluted by human and animal excrement and also by home domestic wastes. The MPN-Coliform density (TPC/ml = 1.1 x 105 and MPN-Coliform = 1.1 x 105 per 100 ml) far exceeded the level recommended for water for public use (MPN-Coliform = 103 per 100ml).

From the data obtained from the brief survey on water quality of the Cimanuk watershed, it could be concluded that the streams of the watershed:

1. were already polluted by suspended solids of medium to severe levels;
2. were already polluted by organic matter (including excement of human and animal origin) of low to medium levels;
3. had Fe and Zn contents, at some places also a Ci content, which exceeded the maximum limit for drinking water;
4. could no longer be used for drinking water or drinking water sources without certain treatments; this is also true of water used for industrial purposes, for cooling, or for water processing; and
5. had a water quality which was found feasible enough for agricultural and fishery uses.

References

APHA 1975. Standard methods for the examination of water and waste water. 14th ed. APHA Inc., New York.

Dent et al. 1977. Detailed reconnaisance land resource survey Cimanuk watershed area.

Pescop, M. B., 1973. Investigation of rational effluent and stream standards for tropical countries. AIT, Bangkok.

Team Kualitas Air Pusdi PSL-IPB and Proyek Pengelolaan Sumbersumber Alam Lingkungan Hidup Panitia Perumus den Rencana Kerja Bagi Pemerintah Di bidang Lingkungan Hidup 1978. A Study on the Determination of Criteria for the Quality of Water and Biotic Environment of the Cimanuk Watershed.

World Health Organization 1963. International standard for water supplies. WHO Monograph, Geneva, 2nd ed.

Discussion

Thayib: Surely the standard required for drinking water should follow WHO standards.

Sutamihardja: True, but the people have to drink whatever water is available.

Punjanan: Is it possible to determine the distance of penetration of sea water into the Cimanuk River by using salinity data?

Sutamihardja: Sampling was only in fresh water, so salinity was not analyzed.

Contents - Previous - Next